Course Outline

🔹 Day 1: Foundations of Responsible AI and Ethical Frameworks

Session 1: What is Responsible AI?

• Definition and importance of Responsible AI
• Overview of responsible AI principles: fairness, transparency, accountability, privacy, and safety
• Key global frameworks: OECD, UNESCO, EU AI Act, NIST AI RMF

Session 2: Ethics in AI Design and Deployment

• The role of ethics in the AI lifecycle
• Value-sensitive design
• Stakeholder impact assessment
• Ethical dilemmas and risk trade-offs

Session 3: Identifying and Understanding Bias

• Types of bias: data, model, labeling, and deployment bias
• Sources and causes of bias in machine learning pipelines
• Fairness metrics: demographic parity, equal opportunity, etc.

Lab Activities:

• Ethical scenario analysis: Group discussion and decision-making
• Bias audit using sample datasets
• Map AI stakeholders and their responsibilities

🔹 Day 2: Explainability and Fairness in Practice

Session 1: Explainable AI (XAI) Concepts

• What is explainability? Why does it matter?
• Post-hoc vs. intrinsic explainability
• Model transparency and interpretability trade-offs

Session 2: Tools for Explainability and Fairness

• Hands-on with LIME, SHAP, and What-If Tool (Google)
• Fairness evaluation with IBM AI Fairness 360
• Visualizing model decisions and feature importance

Session 3: Building Trust in AI Systems

• Human-AI interaction and trust
• Communicating AI decisions to non-technical stakeholders
• Explainability for high-risk domains (health, finance, law enforcement)

Lab Activities:

• Use SHAP or LIME to explain a model prediction
• Evaluate a trained model using AI Fairness 360
• Create an explainability report for a non-technical audience

🔹 Day 3: Governance, Accountability, and Project Work

Session 1: AI Governance and Compliance

• Policies, standards, and auditing AI systems
• Internal governance frameworks (model cards, data sheets)
• Legal compliance: GDPR, AI Act, and industry-specific regulations

Session 2: Designing Ethical AI Systems

• Ethics-by-design principles
• Embedding fairness and accountability in data collection and model design
• Risk mitigation strategies

Session 3: Capstone Project + Presentations

• Develop a responsible AI framework for a use case (e.g., hiring, lending, healthcare)
• Present framework including ethical considerations, mitigation strategies, and explainability
• Group reflection and peer feedback

Lab Activities:

• Build a governance checklist for an AI project
• Analyze a real-world AI failure case and suggest responsible redesign
• Present capstone project and receive peer evaluation